Grants talk:Project/Rapid/Dagbani Wikimedians User Group/Wikidata Lexicographical Data (Abstract Wikipedia and Wikifunctions)

Thanks for submitting this project!

I'm really excited to see community activities around Lexicographical data \o/
Here are two questions to help you streamline your thoughts around the project:

First

Lexemes are only going to be good as useful for Abstract Wikipedia if they have senses, forms, and statements connecting them to other "things".

• Do you have a criterion you will be setting for editors so that the lexemes you will be creating have senses, forms, and statements on them?
• How will you track the 500+ lexemes to ensure that each lexeme has met these criteria?

Second

Another useful way to prepare for Abstract Wikipedia is to translate as many Wikidata Property labels (excluding external identifiers) as possible. You wrote that you'll

add language labels and descriptions to Wikidata properties

• How many labels are currently translated into Dagbani?
• How many more labels will you have translated at the end of the project?

All the best! -—M@sssly^✉ 10:16, 15 January 2022 (UTC)

Hi @Masssly:, Thank you for your questions. Below are answers to your questions:

First

• Do you have a criterion you will be setting for editors so that the lexemes you will be creating have senses, forms, and statements on them?

Ans: Yes. Just like we have done in the past, we will group all 20 participants under each team member (experienced Wikidata contributors) . They will ensure that each Lexeme have senses, forms, and statements on them.

That's a good plan. The majority of the current 120 Dagbani Lexemes are Nouns. It'd be more helpful if there are a variety of different parts of speech other than Nouns. Here is a list of parts of speech, you could have these 20 team leads each focus their group on 1 or 2 parts of speech. In case you're looking for words, Trey made a list of the most commonly used words on Dag Wikipedia. I'm not sure though how up-to-date that list is.

Yes. We would be able to create more Dagbani Lexemes for different parts of speech other than nouns. We are in touch with Trey and thanks for sharing these useful links.

• How will you track the 500+ lexemes to ensure that each lexeme has met these criteria?

Ans: Currently, the Dagbani Wikimedia community have created at least 120 Lexemes in Dagbani. Aside from running a Wikidata query, We have a Google spreadsheet tool that will be made accessible to all 20 participants. This tool will help us track all the 500+ Lexemes and also ensure that each Lexeme have senses, forms and statements.

It's probably better to put your list on wiki. As the lexeme namespace is still pretty niche for new communities like yours, I can also imagine a bit of needed cleanup afterwards. I'd recommend that you create a Listeria tracking page for the project (or manually curate a list) and ping the Lexicographical data community so we can follow your work and provide support along the way.

Good suggestion! Listeria-curated list will be much easier to work with than Google sheet. We will create a Listeria tracking page and reach out to the Lexicographical community if we need help.

Second

• How many labels are currently translated into Dagbani?

Ans: There are currently 84,694 (0.08%) Wikidata labels in Dagbani (See here). Almost all the top 100 Wikidata property labels have been translated into Dagbani (See top 100 list of wikidata properties labels in Dagbani)

\o/

• How many more labels will you have translated at the end of the project?

Ans: At the end of the project, we would have translated al least 100+ Wikidata properties labels into Dagbani.

Great, it looks like there are currently less than 100 properties with labels, but if 100 more property labels can be translated at the end of the project that'd be wonderful.

Yes. The most challenging part of this project is getting Dagbani translations for some property labels. We hope to translate at least 100+ property labels at the end of the project.

Reagards, Shahadusadik (talk) 11:12, 19 January 2022 (UTC)

Cheers! -—M@sssly^✉ 21:34, 23 January 2022 (UTC)

Thanks for your comments suggestions. Shahadusadik (talk) 13:23, 24 January 2022 (UTC)

Comments from DSaroyan (WMF)

Hello Dagbani Wikimedians, thanks for submitting this Rapid Fund request. I have reviewed it and have a few comments:

• We won't fund internet expenses for 25 participants for a full month. We will only fund internet expenses during the actual events. Please remove the first expense from the requested budget.
• I think that 150 GHS per person/day for food is too much. Please reduce this expense and align it with your local prices.
• For pull-up banners, please make only 1 banner and confirm that you will create a reusable banner, so that you can use it for your next similar events. Please make this change to your budget.
• Dear Masssly, thank you for your productive questions. Could you please confirm if most of your concerns are addressed?

Looking forward to hearing from you. Best regards, DSaroyan (WMF) (talk) 16:24, 21 January 2022 (UTC)

Hi @DSaroyan (WMF):,

Thank you for your comments, below are explanations to classify your observations,

• We won't fund internet expenses for 25 participants for a full month. We will only fund internet expenses during the actual events. Please remove the first expense from the requested budget.

Ans: The event will run throughout the month of February. 20 trained volunteers are expected to participate in the month long project. We have reduced the total to 20 participants including team members. The 2 training sessions are meant to train participants on Wikidata items and its lexicographical extensions(Lexemes). We will also equip them with the necessary skills needed to effectively participate in the abstract Wikipedia project (We wont be able to effectively participate without support from the rapid grant as compared to the other 4 selected languages). It is obvious that we can not create over 500+ Lexemes and translate Wikidata property labels from the same training workshops therefore, we are requesting Ghs 500 to rent internet routers and/or data allowance for each participants. This is a huge task considering the amount of time participant will spent in creating the Lexemes and translating Wikidata property labels into Dagbani. I am taking the 5 team members out leaving only 20 community members who will benefit from the internet data/or rented routers. The second internet expense on the budget is for the training sessions and that is for internet data for the training sessions.

• I think that 150 GHS per person/day for food is too much. Please reduce this expense and align it with your local prices.

Ans: For the 2 training sessions, we have reduced the cost to ghs 90 for food and refreshment for each participant.

• For pull-up banners, please make only 1 banner and confirm that you will create a reusable banner, so that you can use it for your next similar events. Please make this change to your budget.

Ans: The pull-up is reduced to 1 as recommended and I confirm that a reusable banner will be created. Warm regards, Din-nani1 (talk) 18:10, 22 January 2022 (UTC)

Yes, they have. I Ieft a few recommendations. —M@sssly^✉ 21:36, 23 January 2022 (UTC)

Hello Din-nani1, thanks for your response. For the internet expenses for individuals, I still think that it should be removed or highly reduced:

• While I understand the need for renting a router for group sessions, we will not fund renting expensive routers for individuals to contribute.
• You mentioned that participants will have their mobile devices to contribute. You may want to offer them data bundles. Since editing Wikidata does not require too much data, I think 100 GHS (or even less) per person would be sufficient for a month if participants use Wikimedia projects only. How much data will you buy with 100 GHS? My quick internet search shows that it should be 8-10 GB data.
• 500+ lexemes means that each participant will edit 20 lexemes in a month-long contest (500/25). It is less than 1 lexeme and ~4 translated Wikidata labels per day during a month. How much data will a user need to create/edit 1 lexeme and translate 4 labels per day? I'm sure less than what you currently request.

So, my request is final. Please remove or highly reduce the internet expenses per individual. If I'm missing anything, please let me know. Best regards, DSaroyan (WMF) (talk) 11:42, 25 January 2022 (UTC)

I have reduced the internet expense per individual as requested. Din-nani1 (talk) 13:10, 25 January 2022 (UTC)